INTRODUCTION
Hepatocellular carcinoma (HCC) is the second most common cause of cancer death worldwide; its risk factors are diverse and evolving, with regional trends in many societies1. In Mexico, mortality due to viral hepatitis, liver tumors, and cirrhosis is increasing in northern and central states2, and HCC was the most common primary liver cancer in 2006 with a mortality rate of 4.7/100,0003. During the early 90s, in Mexico City, a two-fold increase in the incidence of HCC was observed among 12,556 autopsies performed over a 25-year period4. Incidence of cirrhosis was 4% in that sample of middle-income patients, and 14% of HCCs were found in non-cirrhotic livers. In other studies, non-cirrhotic HCC has a frequency that ranges from 1.7% to 19%, in patients without clear associated risk factors5-7. A local study of 1486 cirrhotic patients identified alcoholism (39.5%), hepatitis C virus (HCV) infection (36.6%), primary biliary cholangitis (5.7%), and hepatitis B virus (HBV) infection (5%) as frequent etiologies of end-stage liver disease8, but 154 (10.4%) of them were cryptogenic cirrhosis8.
Tortillas (corn flour flat bread) are a staple food consumed in a daily basis in Mexico, and a recent Mexican study9 showed that 13% contained levels of aflatoxin that exceed the allowed national quantification limits (>12 µg/kg). Aflatoxins are the carcinogenic and mutagenic mycotoxins of Aspergillus parasiticus and Aspergillus flavus. Food contamination mainly with aflatoxins type B1 (AFB1) has been associated with HCC development in Southeast Asia and sub-Saharan Africa10, regions with a high prevalence of HBV exposure. Both are risk factors for hepatocarcinogenesis, but AFB1 induces DNA adducts and consequently promotes oxidative damage and G-T transversion in codon 249 of the TP53 tumor suppressor gene TP5310. The R249S mutation has been found in more than 60% of HCC patients exposed to excess dietary AFB111-13. In fact, two independent studies using polymerase chain reaction (PCR) and mass spectrometry analysis in tumoral and blood samples from regions with a high prevalence of HCC and AFB1 exposure had proposed TP53 R249S mutation as a pre-diagnosis serum biomarker of HCC since 88.5% concordance between tumoral tissue and matched plasma was found13,14. In Guanxi, China, where >50% of HCC had the exposure, a molecular screening of all TP53 exons showed R249S as the most frequent missense mutation in patients with and without HBV infection15. An earlier study analyzing 21 HCC patients from Northeastern Mexico, found 76% with AFB1 serum adducts, and of 16 amplified cases, 3 (19%) had the TP53 R249S mutation16. After this single report, without a confirmatory local study, our country was considered as an intermediate risk zone for this carcinogenic pathway in the liver16,17.
In Latin America, the genotypes H and F of hepatitis B appear to have a more benign course of disease18, with few HCC cases associated with this infection19,20. Our objective was to study whether HCC in Mexican Mestizo patients carry the TP53 R249S mutation linked to aflatoxin exposure and describe its associated risk factors in patients with and without cirrhosis.
METHODS
Data from HCC-treated patients with available liver biopsies, follow-up, and clinical information were identified in surgical pathology files from an academic medical center. Exposure to hepatotropic viruses, alcoholism, metabolic diseases, diabetes mellitus, hypertension, episodes of ascites, portal hypertension, and the body mass index were retrieved from clinical charts. Slides were re-reviewed to identify HCC and remnant hepatic tissue was reassessed for fibrosis and steatohepatitis. Macrodissection of neoplastic cells was performed; DNA was purified from paraffin-embedded tissue samples containing at least 70% of tumor cells using the QIAamp DNA FFPE Tissue kit (Qiagen, Hilden, Germany), following manufacturers recommendations. In all cases, the total amount of DNA obtained was greater than 300 ng. A product of 238 bp that included the exon 7 of TP53 was amplified by PCR. The reaction was performed using 25 ng of tissue DNA, 20 µl of 2× HotStarTaq Plus Master Mix Kit (QIAGEN), and 0.25 µM forward 5´-CTTGCCACAGGTCTCCCCAA-3´ and reverse 5´-AGGGGTCAGAGGCAAGCAGA-3´ primers in a final volume of 40 µl. The PCR reaction consisted of 40 cycles in the following conditions: 98°C for 30 s, 60.8°C for 40 s, and 68°C for 40 s. PCR products were purified using QIAquick Gel Extraction kit (Qiagen, Hilden, Germany) and used as template for direct sequencing with the 3130 × l Genetic Analyzer (Applied Biosystems, Foster City, CA). The sequence NM_000546 (GenBank) was used as a reference. The study was performed in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki).
RESULTS
From the years 2002 to 2015, 74 HCC patients were identified among 1863 consecutive liver biopsies (4%). No clinical data were available in five patients; therefore, they were excluded. Patients submitted to exon 7 screening had a median age of 62 years and 46 (67%) were male. Stage 4 fibrosis was observed in 46 patients (67%); the main risk factors for cirrhosis development were HCV infection (HCV, 39%, 18/46), alcoholism (20%, 9/46), and HBV infection (HBV 2%, 1/46), and in 18 (39%), cirrhosis was recorded as cryptogenic. Stage 3 fibrosis or lower stages were observed in 23 cases (33%) without a demonstrated chronic liver disease. Of these patients, 8/23 had type 2 diabetes mellitus (T2DM) and 6/23, systemic hypertension. Six T2DM patients also had hypertension, two had dyslipidemia, and one had T2DM, hypertension, and overweight (Table 1). Remnant liver tissue was present in 11/23 samples without cirrhosis; in five of these patients, steatohepatitis with interlaminar fibrosis stage 2 was observed. In addition, four patients had tumors with steatosis, abundant Mallory bodies, and satellitosis, one of them also with steatohepatitis in residual tissue. None of them presented the ground-glass cytoplasm nor surface/core antigen expression suggesting HBV infection, as was routinely ruled out by immunohistochemistry.
Risk factor for HCC | Cirrhotic HCC (n = 46) | Non-cirrhotic HCC (n = 23) |
---|---|---|
Hepatitis C virus | 18 | 0 |
Alcohol | 9 | 0 |
Hepatitis B virus | 1 | 0 |
Type 2 diabetes* | 0 | 8 |
Systemic hypertension | 0 | 6 |
Steatohepatitis** | 0 | 5 |
Steatohepatitic HCC | 0 | 4 |
Cryptogenic cirrhosis | 18 | 0 |
*Six patients with type 2 diabetes mellitus (T2DM) also had systemic hypertension, two with dyslipidemia, and one T2DM, hypertension, and overweight.
**One patient had steatohepatitis in remnant liver tissue and steatohepatitic variant of HCC. HCC: hepatocellular carcinoma.
Tumor grading according to the Edmondson-Steiner and WHO modifications21 was well differentiated (G1-G2) 31 (45%); moderate (G3) 23 (33%); and poorly differentiated (G4) 12 (17%), without differences between cirrhotic and non-cirrhotic tissue. Three (4%) fibrolamellar HCC subtypes were observed in livers without fibrosis. A 238-base pair fragment of DNA was obtained in each tumor, and its direct sequencing did not show the AGG-AGT or AGG-AGC mutation expected in codon 249 of the exon 7 of TP53.
DISCUSSION
AFB1 synthesized by Aspergillus spp. is one of the most common carcinogens in foods such as rice, cereals, fruits, vegetables, and corn, among others17. This toxin is chemically stable and resistant to various processes such as cooking or boiling17. Corn contamination with AFB1 was documented early in the 90s in Northeastern Mexico9,16. Recent findings of tortilla contamination with this carcinogen in Mexico City suggest extensive exposure and protean clinical expressions9,16,22. Its most feared complication is HCC development in regions with high prevalence of HBV infection; however, none of our cases in this study carried the TP53 mutation linked to AFB1 exposure. Explanations for this finding could be the low prevalence of HBV as an additional risk factor in the analyzed patients (1 HCC/HBV among 69 HCC). Gain of oncogenic function of TP53 R249S mutation depends on the microenvironmental milieu that HBV creates, leading to post-translational modifications and neoplastic cell survival advantages, as was recently described23. A second explanation could be that AFB1 exposure became inconstant once the liver mass was present, leading to disappearance of its mutational signature through clonal and subclonal evolution of HCC24. Although a previous report states AFB1 exposure as a concern for HCC in Mexico16, our sample with only one case of HBV infection is not confirmatory; instead, it should lead to consider other players in hepatocarcinogenesis8, since hepatotropic viruses and alcoholism merely explained 40.5% of the analyzed cases (28/69).
The frequency of non-cirrhotic HCC in our study was 33%, higher than the previously observed range of 1.7-19%5-7,25. Adding cases of cryptogenic cirrhosis where no known risk factor leading to cirrhosis was found to non-cirrhotic HCC cases, the percentage of HCC without a well-documented risk factor was almost 60% in the present study. Aside of hepatotropic viruses and alcohol, new risk factors have been associated to HCC development26 and should be analyzed in future studies in Mexico, a country where HCC prevalence in postmortem studies rose from 0.35% in 1965-1969 to 0.69% in 1985-19891,4.
Fibrolamellar hepatic carcinoma, the carcinoma subtype usually observed in non-cirrhotic liver27, was identified in three patients, four HCCs were steatohepatitic28, and the rest were conventional HCCs with trabecular and acinar patterns, none of them with the TP53 R249S mutation. Non-tumoral liver tissue showed steatohepatitis in five patients with fibrosis Stage 3 or lower. In a country with high prevalence of obesity and diabetes as is Mexico29, our results suggest that the metabolic syndrome could be involved in patients developing HCC without cirrhosis30.
Although it is important to recognize the limitation of the present retrospective series lacking AFB1-albumin adducts level measurement in serum16,17, in concordance with other studies (Table 2), our negative result reflects the wide variations of this association around the globe31-57. Without diminishing credibility to the previous studies also lacking adduct exploration in serum36,38-40, two studies done in Asia showed lower frequencies of TP53 R249S in that region37,41, and there is the pioneering study of Mexican patients done abroad in a collection of only 16 liver tumors and plasma16. In our setting, where end-stage liver disease screening is exhaustive, steatohepatitis and conditions associated to metabolic syndrome warrant the reassessment of clinical records and liver tissue beyond the well-known classical risk factors for hepatocarcinogenesis. The silent natural history of liver lipotoxicity is starting to be unveiled in Mexico58, and more and larger studies on chronic liver damage, with in-depth clinical and molecular analyses, are required to elucidate further our interesting findings.
Region | Specimen | TP53 R249S+ve/HCC (%) | Detection method | R249S mutation and serum AFB1 adducts | Year | Reference |
---|---|---|---|---|---|---|
Africa | ||||||
Southern Africa | Tumor | 3/10 (33) | Sequencing | ND | 1991 | 31 |
Mozambique | Tumor | 8/15 (53) | RFLP | ND | 1991 | 12 |
Senegal | Tumor | 10/15 (67) | RFLP | ND | 1993 | 32 |
Gambia | Tumor | 11/18 (61) | RFLP and sequencing | ND | 2004 | 14 |
Gambia | Tumor | 43/56 (77) | RFLP | ND | 2012 | 33 |
Nigeria | Plasma | 6 (7) | RFLP and sequencing | ND | 2008 | 34 |
Egypt | Tumor Serum |
2/20 (10) 1/76 (1) |
RFLP and sequencing | ND ND |
2008 | 35 |
Asia | ||||||
Qidong, China | Tumor Plasma | 11/18 (61) 2/130 (2) | RFLP SOMA | ND ND | 2009 | 36 |
Guangxi, China | Tumor | 18/50 (36) | Sequencing | ND | 2001 | 37 |
Beijing, China | Tumor | 1/116 (1) | PCR-SSCP and sequencing | ND | 2013 | 38 |
Thailand | Tumor | 6/25 (24) | SOMA | ND | 2005 | 39 |
Thailand | Plasma | 29/84 (35) | SOMA | ND | 2012 | 40 |
Japan | Tumor | 3/279 (9) | Sequencing | 3/3 | 2011 | 41 |
Japan | Tumor | 3/34 (9) | PCR-SSCP and sequencing | ND | 1993 | 42 |
Japan | Tumor | 7/140 (5) | PCR-SSCP and sequencing | ND | 1992 | 43 |
Japan | Tumor | 0/59 (0) | PCR and OH | ND | 1993 | 44 |
Singapore | Tumor | 0/44 (0) | RFLP | ND | 1995 | 45 |
India | Tumor | 2/21 (10) | PCR-SSCP and sequencing | ND | 2000 | 46 |
Europe | ||||||
Italy and France | Tumor | 0/7 (0) | Sequencing | ND | 1993 | 47 |
Great Britain | Tumor | 0/19 (0) | PCR-SSCP and sequencing | ND | 1992 | 48 |
Germany | Tumor | 0/20 (0) | Sequencing | ND | 1995 | 49 |
Italy | Tumor | 0/20 (0) | Sequencing | ND | 1995 | 50 |
Turkey | Tumor | 1/50 (2) | RFLP | ND | 2010 | 51 |
America | ||||||
Brazil | Tumor | 21/74 (28) | RFLP | ND | 2009 | 52 |
United States | Tumor | 4/37 (11) | RFLP | ND | 1993 | 53 |
Hispanics in South Texas | Tumor | 3 /41 (7) | RFLP | ND | 2018 | 54 |
Alaska | Tumor | 0/13 (0) | Sequencing | ND | 1995 | 55 |
Colombia | Tumor | 4/38 (11) | RFLP and sequencing | 0/4 | 2011 | 56 |
Colombia | Tumor | 1/30 (3) | RFLP | ND | 2019 | 57 |
México | Tumor | 3/16 (19) | Sequencing | 2/3 | 1996 | 16 |
Mexico | Tumor | 0/69 (0) | Sequencing | ND | 2020 | Present study |
RFLP: restriction fragment length polymorphism; PCR-SSCP: polymerase chain reaction and single-strand conformation polymorphism; OH: oligonucleotide hybridization; SOMA: short oligonucleotide mass analysis; ND: not determined.
In the absence of HBV infection, no evidence of aflatoxin involvement in hepatocarcinogenesis was revealed in this retrospective study. More than half of HCC lacked a known risk factor for end-stage liver disease. In addition to the traditional risk factors, steatohepatitis or conditions associated with metabolic syndrome were frequent in 69/1863 (4%) HCC obtained from consecutive liver biopsies.